At the same time, we saw models bit by bit, speed up the timing of the changeover to sleet. Of course this had mostly go do with the storm slowing down and the ULL cutting west of the Apps. I do think in general, the models had a reasonable grasp on the temp profiles in the end. Point is, I think most of the time outcomes have more to do with the track of the low, strength of the high, etc. than bad physics of the model